In collaboration with the Centre Regional de Medecine Traditionnelle in Bandiagara Mali, this longitudinal project investigates how migration for work and other socio-environmental factors affect psychological functioning, mental and physical health as well as the nature of AIDS-related attitudes and behaviors, in the Dogon (primarily farmers), Fulani (primarily herders) and Bozo (primarily fishermen) ethnic groups. It involves an interview of a representative sample of 1002 rural Malians. The three-pronged protocol reflects the two units' joint and individual concerns. One prong is the test of the generalizability to rural Mali of the SSES findings in industrially developed societies that doing self-directed and substantively complex work leads to more self-directed and autonomous orientations and more effective intellectual functioning. The second prong is an examination of the consequences for mental health of stressful occupational conditions, including migration for work. The third prong is the examination of the effects of a) migration for work and b) cultural and socioenvironmental factors on AIDS-related knowledge, attitudes and behaviors. The first wave of data collection was conducted in 1996/1997, the second wave in 2003/2004. [unreadable] [unreadable] This year, using the data from both waves, we developed a series of structural equation measurement models of key concepts (e.g., substantive complexity of work, authoritarianism, self-esteem, anxiety/depression) that parallel those we have developed using parallel data from similar surveys in industrialized countries. To our surprise, the overtime correlation over the six year interval of almost every one of the psychological variables was notably much lower than had been found in any other cultural setting. In order to ensure that these low over-time correlations were not a result of some mistake of ours or statistical artifact, we went through the laborious processes of: 1) fully rechecking the data entry, 2) extensively reestimating models representing each of the psychological variables involved. When we did so, we found nothing that cast doubt on the validity of our initial findings. The very low level of statistical stability of these important psychological variables in the population we are studying is real and not artifactual. We are presently testing the hypothesis that this real psychological changeability in our populations is the result of dramatic societal change.